Nvidia says its inference accelerator Groq 3 LPX has entered full production and Nebius has signed on as the first customer; SpaceXAI will adopt Vera CPUs (Mike Wheatley/SiliconANGLE)